New AI uses real-time weather insights to cut data center cooling costs by 25%

Penn State researchers developed AI software that cuts data center cooling costs by 25% using real-time weather and electricity price data. The system replaces static cooling with adaptive strategies, dynamically adjusting temperatures based on external conditions while maintaining safe operating limits. Since cooling accounts for 40% of data center electricity use, this physics-based AI model offers significant savings without requiring expensive hardware upgrades.